Constructing Gene Co-expression Networks for Prognosis of Lung Adenocarcinoma

Many studies of prognostic genes for cancer have focused on comparative analysis of gene expressions in cancer cells and normal cells. However, prognosis of cancer patients can be done more accurately by comparative analysis of patients with different conditions. In this study we partitioned the patients with lung adenocarcinoma into two groups, one with a wide-type TP53 gene and the other with somatic mutations in the TP53 gene, and constructed gene co-expression networks for the two groups. From the comparative analysis of the two GCNs we obtained several gene pairs with significantly different co-expression patterns in the two groups. The GCNs constructed in our study are more informative than other GCNs in the sense that ours provide the specific type of correlation between genes, the concordance and prognostic type of a gene. The GCNs will be informative for prognosis of lung adenocarcinoma, which is the most common type of lung cancer.